Gait pattern-aware displacement estimation in micro-sensor motion capture

Micro-sensor motion capture (MMocap) systems do not have external references. It is challenging to estimate the displacement of human CoM (Center of Mass). This is especially true when gait pattern changes. This paper presents a novel gait pattern-aware algorithm that aims to estimate the displacement of human CoM during walking and jumping for MMocap systems. Firstly, we fuse the acceleration of sensors and hierarchical information of the body to detect the gait patterns. Secondly, according to the detected patterns, the displacement is adaptively calculated. 1) During walking the displacement is calculated by the extension and flexion angles of the lower body joints. 2) During jumping the displacement is calculated by integrating the acceleration of the root joint. The experimental results have demonstrated that our algorithm estimates the displacements accurately during both walking and jumping.